Category: Articles
Hits: 2044


Renat Terlecki,DC power amplifier, Antena 1934/11

Radio engineering has recently gained a completely new field of activity in the form of high power loudspeaker installations. Apart from the film of high-power loudspeaker installations, we also often meet in a wide range of public utilities such as train stations, guest houses, hotels, sports grounds, concert halls, even churches and parliaments.
Several years of practice crystallized the concept of a modern power amplifier and today we require from such a device:

  1. Complete electrification.
  2. Uniform gain and reproduction of a wide frequency range.
  3. Versatile use, i.e. for both the adapter and the microphone, photocells, radio and so on.
  4. High efficiency possible.
  5. Transparent design.
  6. Simple operation and minimal maintenance, and finally
  7. Carefully fitted equipment.

However, all this is successful when we have an AC source available, usually in the form of a lighting network. Alternating current, as we know, can be transformed into any high or low voltage; so we can build economical filament circuits without losses on reducing resistors, and having any high anode voltage we use more efficient systems, eg "class C", i.e. directly coupled.

In many cities, however, we have power plants, supplying lighting or industrial DC, and in this case building a high power amplifier, we face special difficulties. In fact, for the reasons given above, high power amplifiers are usually calculated on the power network. variable, so wanting to power them from the power grid fixed, we must use converters whose cost for small installations may exceed the value of the amplifier. So sometimes it pays to give up the economics of the system and settle for a relatively high power amplifier with a very low anode voltage of 150 ÷ ​​200 Volt.
The amplifier described below is designed for a 250 ÷ 150 Volt DC power supply. Understandably, by reducing the value of the main reduction resistance "R" we can connect the amplifier to a network with a lower voltage, but then the output stage power would drop too much.

SCHEMATIC DIAGRAM

Before we get to our amplifier, let's start by studying the most important system - the filament circuit. This circuit, as we can see from the simplified diagram in Fig. 1, consists of adjustable resistance R5, at whose ends, according to Ohm's law, the flowing current (1.4 Amp.) Creates the potential difference "Eg", used here as the negative pre-grid of output tubes. Then the filament current branches off after 0.65 Amp. on two cathodes of output tubess in parallel. The rest of the current, according to Kirchhoff's law, flows through the regulated shunt resistance R4, which regulates the filament voltage of both tubess. Then again, the sum of the filament current flows through the heater of the indirectly-heated input tube (ca 1 Amp.), whose heater is controlled by the R3 shunt resistance. Next we have an iron - hydrogen tube "1331" automatically regulating the filament current to 1.3A - and finally the main reduction resistance R.

 

Fig.1

It should be noted that the sum of resistance of this circuit between points A and B must be a constant value for a given network voltage. Not very large fluctuations in the network voltage and changes in R5 and R resistances will automatically compensate the resistance of 1331 tube, however, e.g. larger changes in R5 resistance must be compensated with R resistor. As for the essential part of the amplifier, as we can see from the complete diagram in Fig. 2, it is very simple.

Fig.2

At the input we have a switch that allows you to quickly transfer the amplifier from one pair of input sockets to another. From the switch, the current of audible frequency runs through the Pot potentiometer, regulating the force, from where it goes to the primary winding of the transformer Tr1, whose secondary winding is the grid circuit of the amplifier's input tube. The potentiometer can be successfully lowered if the audible current source (adapter-microphone) used has its own power regulation. The resistance R2 inserted into the cathode circuit of the first tube gives it the necessary negative pre-netting. The input tube anode through the primary winding Tr2 is connected to the choke Dl, which together with the C2 capacitor is used to smooth the pulsation of the network current. The Tr2 secondary winding has three terminals.

Negative grid voltage of the output tubes is connected to the center of this winding, while the other terminals, on which we receive alternating voltages of amplified currents with a phase shift of 1800 are connected respectively to the grids of these tubes. From the above, we can see that the output tubes work here in a symmetrical anti-push system, so-called "Class A". The anode currents of these tubes, or rather the variable components of these currents, add up in the "Transformer" output transformer, whose primary winding is also divided into two sections, attached to the anodes of both tubes in opposite directions. Since the phases of these currents are also opposite, so the magnetic fluxes generated by them in the core have the same directions, and therefore they add up. On the secondary side, the output transformer has 2 windings: for a low and high-resistance loudspeaker with branches, which again allows the transformer to be adapted to the loudspeaker.

The above described system of output tubes is known to have exceptional uniformity of gain, and most importantly, it does not require high anode voltages. In the above considerations, we omitted R1-C1 and R6-C3, which, as shown in the diagram, are used to filter individual grid voltages; it should be noted that the anode voltage of the end tubes is not filtered at all here, because its pulsations cancel each other out on the output transformer, and even bypassing the resistance R6 and directly connecting the center of the secondary winding Tr2 to the minus of the network is simply not felt.

LIST OF PARTS

Pot = 50000ohm; Transformer Tr1 = intertube 1: 3 - 1: 6, high class; Tr2 = Polton, Pusch-Pull, heavier 1: 4; Tr. w. = output Polton Pusch-Pull, type WDM3 with winding for dynamic and magnetic loudspeakers; mA = milliammeter with 0-200mA moving coil; Length = Polton D3530 cable gland; B = 3.5V-2A bulb fuse. Capacitors: C1 = C3 = 2µF; C2 = 4µF 1000V; C4 = C5 = 2 = 4µF. Resistances: R1 = 0.1 meg; R2 = 1000 ohm wire; R3 = 20 ohm variable; R4 = 30 ohm variable; R6 = 0.01 meg; R5 = 20 ohms as described; R = 150 ohms as described. W1 = 2-pole rotary switch; W2 = 1-pole mains switch; 2 4-leg tube stands; 1 5-foot base; 1 3-foot base; 2x3 switch; 6 threaded rods, 4mm each 14cm long with 48 nuts; asbestos sheet, 320x250mm asbestos sheet, 400x420x10mm plywood; bakelite 400x110x3mm; 8 telephone sockets on two bakelite plates, 11 5mm screws with nuts for chassis.

In the model amplifier, I used the following tubes with excellent results: "AR4101" in the input stage and in the output stage 2 x "P460" (Tungsram) as a network voltage equalizer, and actually the regulator of the glow current of the tube "1331" (Philips).

ASSEMBLY

We start by making a chassis (Figs. 3, 4 and 5). From a 400x300x10mm piece of plywood, we cut a 400x170 mounting board with a projection in the left rear corner of 130x70 and put them on two frames made of thick iron sheet. The remaining part of the plywood is laid out on one side with asbestos and set perpendicular to the mounting board. In this way we get a complicated chassis, because in the front part we have "two floors", and in the back there is a place "on the floor" for a resistance tube and, well cooled and thermally insulated from the proper amplifier, space for R and R5 resistors. We will fasten these spaces using the threaded rods and metal strips as shown in the drawings.

Fig. 3

Fig.4

Fig. 5

We will have to make the abovementioned resistors R and R5 ourselves. We cut seven bars 340x50mm from the eternit tile and drill two holes for rods to fix them on the rods and two for the ends of the resistance wire. Then we wrap 10 meters of 0.5mm nickel wire on one such strip - it will be R5 resistance. We wind 14 meters of the same wire on the remaining six. After joining them in series we get resistance R. When winding, the wire should be stretched strongly so that it does not slip during heating. Then we make two buckles from copper sheet, with which we will adjust the resistances R and R5.

We mount a switch, potentiometer (if used) on the bakelite faceplate from the left, in the center a milliammeter, then a fuse and switches W1 and W2.

Place all transformers and tube's stands on the mounting board. Mount the base for the resistance tube on the special projection of this board in the left corner. The rest of the parts will fit underneath.

Installation will be made with 1.5mm2 wire in rubber insulation, while the wires shielded in Fig. 2 - with a covered cable. Cable sleeves and solid capacitor boxes are connected together and possibly grounded. Grounding the transformer cores at the network with a grounded plus gave a negative result.

After installing the amplifier and checking the connections according to the diagram in Fig. 2, we connect it to the network. However, it is not recommended to use a cord with a plug here, because it will be easy to confuse the poles, although there is no risk of damage to the apparatus, it is better, however, after recognizing the poles of the network, e.g. with a polar voltmeter, connect it permanently directly to the W1 switch. Because it is a bipolar switch, after switching off the power the amplifier is completely disconnected from the mains and there is no fear of shock, which is again particularly important at the first adjustment.

FIRST TRY

Before turning on the current, we additionally insert an ammeter on the scale with a scale that would allow us to read 1.4 Amp; then set the resistor R3 and R4 sliders on half of the windings, set the slider on the resistance R5 so that about 3/4 of its resistance wire is turned on, finally the slider on the resistance R set so that all resistance is turned on. Now we set the tubes and making sure that W2 is off; turn on the network by turning the W1 switch. The pointer of the ammeter will move to a certain position indicating approximately 1 Amp. Now we look at the tubes of the P460 tube or by chance the cathode of which does not glow bright red heat. Should this happen, immediately turn off W1 and look for a break in the filament circuit of the second tube, because the cathodes of these tubes are not lit during operation, but if one is lit it would be proof that all the filament current (and in this case about 1 Amp) flows only through this one glowing tube. After removing any possible damage from the railway, we can proceed to the proper adjustment and therefore turn on the power (W1).

Suppose we read 1.1 Amp from an ammeter. We turn off the network and move the clamp on the resistance R so as to slightly reduce its value and turn on W1 again and read the value of the current flowing from the ammeter. We proceed in this way until 1.3 Amp is obtained. This will be the correct amount of current flowing in the filament circuit (Fig. 1).

Now, by moving the resistance slider R3 we set it so that the voltmeter attached to the legs of the glow of the indirectly heated tube indicates 4 volts. In a similar way, by adjusting the resistance of R4 we set the glow of "P460" tubes to 3.5 Volts.

After completing this operation, we can turn on the anode voltage of the tubes with the W2 switch, while watching the "P460" tubes. If now one of these tubes lights up with bluish or violet light in the inter-electrode space, then W2 should be turned off immediately, then W1 and the lit tube should be sent back to the supplier with a complaint, stating in the complaint card that there was gas in the tube. This is a serious matter because with a strong gas and too large fuse B, the milliammeter and the corresponding half of the output transformer winding may burn, and finally the tube itself will be destroyed due to the rapid cathode bombardment.

This phenomenon, however, should not be equated with another, harmless symptom: glass fluorescence. Sometimes, in tubes with more power during operation, the balloon glass is covered on the inside with a layer of vibrating, delicate celadon light, vividly resembling the X-ray fluorescence of the screen.

So if there was no gas in the tubes, then we read the anode current from the "mA" milliammeter. For a 220V network, it should be around 100mA, i.e. 50mA per tube. If the milliammeter shows more, then you need to increase the value of resistance R5, moving the clamp on its winding and vice versa if it shows less - it should be reduced. After this action, the ammeter will show us 1.4 Amp. as the sum of the glow current = 1.3A and the anode current of the output tubes = 0.1A.

The correct value of the anode current indicated by the milliammeter can be easily calculated from the admission power of the tube. The admission power of the "P460" tube is 10 Watt, so of the two we have 20W. Suppose that the voltage on the anodes of these tubes (not the mains!) measured with a good voltmeter is 200 Volt. So we get 20W: 200V = 0.1A = 100mA. At other network voltages, the anode voltage will naturally change, and the anode current will naturally change accordingly.

Finally, connect the voltmeter to the tube legs "1331" and adjust the resistance R so that the read voltage is about 10V; we adjust the voltage of tubes incandescence with R3 and R4 resistance when W2 to 1 Volt is on and turn off the now unnecessary ammeter.

OPERATION

We connect the adapter cables to one pair of input sockets or directly to a pair of contacts in a switch. To the second - we can attach a second adapter or some other device, e.g. a preamplifier from the photocell (in the cinema) or anode circuit of the receiver lamp and the like. A dynamic or magnetic loudspeaker, or both at the same time connect to the appropriate sockets of the output transformer. If your transformer does not have a winding for a magnetic loudspeaker, it can be connected through two block capacitors of 2 ÷ 4µF to the ends of the primary winding of this transformer according to Fig. 2 (C4-C5); then the transformer plays the role of a low-frequency reactor

Then turn on the network by turning the W1 switch and wait about 1 minute until the cathode of the indirectly heated tube warms up, because the P460 lamps will warm up earlier; after this time has elapsed, turn on the anode voltage with the W2 switch - the milliammeter arrow will jump to the value set above and the amplifier is ready for use.

Actually, the W2 switch could be omitted, because it complicates the service, but if you do not care about simplifying it and you attach great importance to the durability of the tubes, its use is recommendable.

After setting the plate and releasing the adapter, set the force regulator to max. and we observe the milliammeter arrow: if it will go back towards zero with stronger sounds, it means that we gave too little mesh foreposition and vice versa if we get the arrow deflection to max. - this finger is too big. We correct this error by changing the resistance R5. When properly adjusted, the arrow will stay in place during the amplifier's operation.

In the above arrangement, the directly-heated output tubes are DC-heated, it follows that the negative cathode ends of these tubes will be much more heavily loaded than the positive ones, so it is recommended to evenly wear them at intervals of several days (e.g. 2 times a week) the direction of the glow current through the tubes. It is best to take this circumstance into account when mounting the amplifier and bring the filament current to the tube bases so that the adjustment of both tubes gives this change, so we pass the current through one tube, e.g. from the left leg to the right, and through the other from the right to the left. The above has been included in the assembly diagram.

The above amplifier, after nearly a year of operation, in very difficult local conditions, did not show any damage.

Adapters for the described amplifier should be used of high resistance type, because when using low resistance types, the input transformer ratio should be changed to 1: 10 ÷ 1: 20. We take discs with an electric recording, because they give the natural sound of the voice and the so-called the roundness of the throne, which can not be said about discs recorded in the old way. It should be emphasized that for some discs recorded in a special way, such as "Pathé" or "Edisson", where the needle makes vibrations not horizontal but vertical, adapters on the market are not suitable at all.

Renat Terlecki